Publication

Abstract : The availability of new information can often necessitate rerouting of a vehicle. In [1], the set of nonoverlapping useful vicinities of the reconnaissance sites, S, and the 3D space over them was represented using a discrete graph-model. Here, we consider two cases in which an unmanned reconnaissance aerial vehicle must be rerouted. The first scenario involves not visiting useful vicinities, S t ⊂ S, due to newly reported unacceptable threatlevels in their neighborhood. In the second scenario, visitations to useful vicinities, S r ⊂ S, is made redundant due to decreased importance of the sites to which they offer a vantage observation-point. Our earlier graph-model [1] is general enough to provide a basic framework for consideration of rerouting in both cases. We give an algorithm to obtain a graph-model, from the earlier one, that allows efficient gene encoding and efficient application of GA operators to obtain new alternative rerouting plans.